MATH 365 MEASURE THEORY AND PROBABILITY
MEASURE THEORY AND PROBABILITY
MATH 365
DEPARTMENT: Mathematical Sciences
MEASURE THEORY AND PROBABILITY
Time allowed: See instructions on Vital
INSTRUCTIONS TO CANDIDATES: This assessment carries 50% towards the
overall module mark. The paper contains 6 questions. Full marks will be given
for complete answers to four questions. Only the best four answers will be taken
into account.
By submitting solutions to this assessment you affirm that you have read and
understood the Academic Integrity Policy detailed in Appendix L of the Code of
Practice on Assessment.
Paper Code MATH 365 Page 1 of 3 CONTINUED
1. Consider the probability space (Ω = [−1,+1],F = B[−1,+1], P = 12m).
(i) Construct the σ-field G generated by the event A = [0, 1]. [7 marks]
(ii) Consider the random variable
X =
{
1, if ω ∈ [0, 1];
0, if ω ∈ [−1, 0)
and construct the σ-field FX generated by X. Are the σ-fields G and FX inde-
pendent? Justify your answer. [10 marks]
(iii) Calculate E(X|G). [8 marks]
2. Let Ω1 = Ω2 = [0, 1] and consider the following function on Ω1 × Ω2:
f(x, y) =
1
x3
, if 0 < y < x < 1;
− 1
2y3
, if 0 < x < y < 1;
0 otherwise.
(i) Calculate
∫
Ω2
(∫
Ω1
f(x, y) dx
)
dy. [10 marks]
(ii) Calculate
∫
Ω1
(∫
Ω2
f(x, y) dy
)
dx. [10 marks]
(iii) Does the statement of the Fubini Theorem hold? Give explanations.
[5 marks]
3. (a) Prove that, for any two sets, A = B if and only if A4B = ∅.
[10 marks]
(b) For the denumerable sequence of sets A1, A2, . . ., we define the sets E and
F as follows
E =
∞⋃
n=1
∞⋂
m=n
Am; F =
∞⋂
n=1
∞⋃
m=n
Am.
(i) For particular case of A1 = {0}, A2 = {0, 1}, A3 = {0}, A4 = {0, 1},
A5 = {0}, A6 = {0, 1}, . . ., calculate E and F . [5 marks]
(ii) Prove that, in general, for any n ∈ N,
En =
∞⋂
m=n
Am ⊂
∞⋂
N=n
∞⋃
m=N
Am.
[10 marks]
Paper Code MATH 365 Page 2 of 3 CONTINUED
4. Let (Ω,F , P ) be a probability space and let ξt be a sequence of mutually
independent random variables on it with
∫
R ξt(ω)P (dω) = 0 and
∫
R ξ
2
t (ω)P (dω) =
σ2t < ∞ for all t = 1, 2, . . .. Let {Ft} be the natural filtration for the process
ξt(ω) and consider the random process
Xt =
(
t∑
i=1
ξi
)2
−
t∑
i=1
σ2i .
Is X an {Ft}-martingale, or sub-(super-)martingale. Justify your answer.
[25 marks]
5. Consider the measurable space (Ω = [0,∞),B[0,∞)). Let Fn be the minimal
σ-field containing all the intervals [0, 1), [1, 2), . . . , [n− 1, n), n = 1, 2, . . ..
(i) What kind of subsets of Ω are elements of Fn ? [15 marks]
(ii) Is {Fn, n ≥ 1} a filtration? Justify your answer. [10 marks]
6. Suppose Ω = (0, 1), F is the σ-field of Lebesgue-measurable subsets of Ω,
and m is the Lebesgue measure. Consider the following sequence of functions on
Ω:
f1(x) = 1;
f2(x) =
{
2, if x ∈ (0, 1
2
);
0, if x ∈ [1
2
, 1);
f3(x) =
{
0, if x ∈ (0, 1
2
];
2, if x ∈ (1
2
, 1),
and so on:
f4(x) equals four only for x ∈ (0, 14); otherwise it is zero,
f5(x) = 4 only for x ∈ (14 , 24), f6(x) = 4 only for x ∈ (24 , 34), f7(x) = 4 only for
x ∈ (3
4
, 1),...
if i = 2n, 2n + 1, . . . , 2n+1 − 1 then
fi(x) =
{
2n, if x ∈ ( i−2n
2n
, i+1−2
n
2n
);
0, otherwise.
Here n = 0, 1, 2, . . .
(i) For an arbitrary x ∈ Ω, compute lim supi→∞ fi(x) and lim infi→∞ fi(x).
[13 marks]
Are the following statements correct?
(ii) limi→∞ fi(x) = 0 almost everywhere on Ω. [4 marks]
(iii) fi converges to zero (as i→∞) in measure. [4 marks]
(iv) fi converges to zero (as i→∞) in Lp-norm, p = 1, 2, . . . . [4 marks]
In all these cases, you should justify your answer.